Major bacterial isolate and antibiotic resistance from routine clinical samples in Southern Ethiopia

Currently, antibiotic-resistant bacterial infections are a challenge for the health care system. Although physicians demand timely drug resistance data to guide empirical treatment, local data is rather scarce. Hence, this study performed a retrospective analysis of microbiological findings at the Hawassa public hospital. Secondary data were retrieved to assess the prevalence and level of drug resistance for the most common bacterial isolates from clinical samples processed at Hawassa University Comprehensive Specialized Hospital. Out of 1085 clinical samples processed in the microbiology laboratory, the prevalence of bacterial infection was 32.6%. Bacterial bloodstream infection was higher in children than in adults (OR, 4; 95% CI 1.8–14.6; p = 0.005). E. coli and K. pneumoniae were the commonest bacterial isolate both in children (36.8%, 26.3%) and in adults (33.3%, 26.7%) from the urine sample while, the leading bacteria identified from the CSF sample was P. aeruginosa, 37% in children and 43% in adult. In this study, all identified bacterial isolates were multi-drug resistant (MDR) ranging from 50 to 91%. The highest proportion of MDR was S. aureus 91.1 followed by K. pneumoniae 87.6%. Since the nationwide investigation of bacterial isolate, and drug resistance is rare in Ethiopia, a report from such type of local surveillance is highly useful to guide empirical therapy by providing awareness on the level resistance of isolates.


Methods
A retrospective cross-sectional study was conducted at the microbiology laboratory of HUCSH, Hawassa, Ethiopia. HU-CSH is a teaching hospital providing services to more than 5 million people in the region as well from nearby region. Likewise, the HU-CSH microbiology laboratory also responsible for bacterial culture and resistance tests for patients in the area. The laboratory was supervised and supported by Federal government through the Ethiopian public health institute (EPHI). The EPHI makes mentor and monthly technical support including external quality assessment.
Samples were collected by the physician or the nurses from all inpatient and outpatient departments and immediately sent to the microbiology laboratory. The decision to take samples for microbiological culture and the selection of samples was made by the physicians. While, patient-related data (age and sex) with a full record of bacteriological culture and antimicrobial resistance profile were retrieved from the laboratory registration book. The study retrieved all microbiological reports on bacterial pathogens from January 2019 to December 2020.
Study participants and data collection. The study participants were individuals who visit HU-CSH and had complain of any infection suspected for bacterial infections during the study period. This study retrieved all documented data's including age and sex of patients, the bacteria isolated and the drug susceptibility profiles from HU-CSH microbiology unit registration books using a standard data collection form.
All laboratory procedures were performed in accordance with the relevant guidelines and regulations of the Clinical and Laboratory Standards Institute's (CLSI) 32 and accustomed to the local standard operation procedures (SOPs) of our microbiology unit. Data obtained in the course of the study were kept confidential and used only Quality control. The quality of our laboratory culture system was performed using E. coli (ATCC-25922), P.
aeruginosa (ATCC-27853), and S. aureus (ATCC-25923) as reference strains for culture and susceptibility testing. Thus, qualities of prepared media were monitored all the time according to the set criteria by CLSI in fact EPHI closely supervised our laboratory monthly.
Data analysis. The data was entered and analysed using SPSS version 20 (IBM Corporation, Armonk, NY, USA). Descriptive statistics like frequency and percentages of categorical variables were calculated. Bivariate comparisons using Chi-square logistic regression were employed to assess the association between variables. A p-value of less than 0.05 was considered statistical significant.

Ethical clearance. Ethical clearance was obtained from the institutional review board (IRB) of Hawassa
University College of medicine and health science. As the study was retrospective, following IRB clearance, we have asked the hospital administration for a waiver to retrieve the data. Accordingly, the official waiver letter was obtained from the HU-CSH hospital administration. We have also obtained official permission from the hospital laboratory manager of HU-CSH. All data obtained in the course of the study were reserved confidential and used only for this study.

Result
Study population. In this study, a total of 1085 clinical samples were processed. Of these, females were onethird 385 (35.5%) of the study population. Likewise, the rate of positivity in female is lower 114 (32.2%) compared to males 240 (67.8%) but not significantly associated (p = 0.11). Regarding the age group, the proportion of under-5 children was about two-thirds (61.3%) of participants. There was a significant association between age group of participants and positivity of bacterial infection (p = 0.154). The age range of participants was from zero-day neonates to 82 years old elders with a mean age of 8.7 years (Table 1).
Of 1085 clinical sample diagnosed in microbiology laboratory 32.6% (354/1085) of them were positive for any of bacterial infection. The proportion of positivity was decrease with age (Table 1).  (Table 3). Nearly 46% (n = 162) of the isolate were attained from patients those who have one or two types of the previous history of antibiotic exposure before specimen collection. The children were more positive in bloodstream infection than adults (OR, 4; 95% CI 1.8-14.6; p = 0.005) ( Table 2). The isolates from different groups of infections were not similar however, common bacterial pathogens were documented in both adults and children (Tables 2, 3). For instance, nearly equal proportions of E. coli and K. pneumoniae isolates were observed in children (36.8%, 26.3%) and adults (33.3%, 26.7%) from urine samples. Again, a comparable result was recorded for P. aeruginosa in the CSF sample, 37% in children and 43% in adults. On the other hand, the most frequent isolate from blood cultures in children was CoNS (29.5%) followed by K. pneumoniae (25.7%) and S. aureus (20%). But in adult CoNS (60%) followed by S. aureus (20%) were found (Table 3). Table 4, most of the gram positive isolates showed the highest resistance to erythromycin and gentamicin for instance; S. pyogens (75%), S. aureus (55%). CoNS showed about 30-55% resistances to Trimethoprim-Sulphametoxazole, ceftazidime, erythromycin and gentamicin. The Enterococci spp. also showed 85.7% resistance to gentamicin and 57.2% to Trimethoprim-Sulphametoxazole, respectively.

Antibiotic resistance. As indicated in
Among gram-negative bacteria, K. pneumoniae has shown the highest resistance rates to ampicillin (75%), ceftazidime (82%) and ciprofloxacin (80.9) while the least resistance (3.3%) to meropenem was documented. Likewise, E. coli also showed the highest resistance to cotrimoxazole (70.4%) and ciprofloxacin (63.6%) while less resistant to meropenem (4.5%). In general, in this study gram-negative isolates were susceptible to meropenem (Table 5).

Pathogen (n)
Antimicrobial   www.nature.com/scientificreports/ were the major isolates in bloodstream infection [43][44][45] . In our finding, P. aeruginosa and K. pneumoniae are the major bacteria isolated from CSF unlike others report 5,38,41,46 . On the other hand, study from Tanzania has reported that S. aureus was the main pathogen in bloodstream infection unlike our finding 47 . Based on this finding, CoNS was the major pathogen in bloodstream infection. The possible attribute might be the poor hygienic care as well as the higher nosocomial infection in our setup 48 that significantly contributes to the high proportion of CoNS. A recent report also showed the highest prevalence of hospital-acquired bacterial infections in a developing country 49 .
In regard to the isolates from UTI, E. coli and K. pneumoniae were documented as major bacterial agent which is consistent with a study conducted with others 41,50 . Our finding was slightly higher than previously published results in Ethiopia 35,36 . Of course, the variation in bacterial isolate of UTI infection has been seen across region 50 . This could be due to differences among study participants, catheterization and hospitalization history.
Unlike our previous report, where N. meningitis and E. coli were the predominant isolate in CSF 38 , in this study P. aeruginosa and K. pneumoniae were the leading isolate. In this study the highest number of bacterial infection were documented in neonates which shows that newborns might be at higher risk of bacterial infection. Hence adequate attention should be given particularly in maintaining the proper hygiene at the ICU 51 .
From other specimens like stool, pus, sputum and discharges, S. aureus and K. pneumoniae were found a major infectious agent. Compiling such type of laboratory database is useful to provide valuable insight about the types of sample and identified drug resistant isolates in areas where microbiological investigations are scarce 41 . Hence our data can be used in support of the existing empirical therapy. Now a day, MDR bacterial infections due to ESKAPE groups become a major health problem and account for a great burden in the effective therapeutic strategies 52 . Our study showed high-level ESKAPE groups MDR (greater than 70%) which is comparable with others report in Ethiopia 41,42 . However, certain variation in bacterial isolate and resistance level reported within-country 5,39 besides related reports in Africa 43,45 . This might be due to a lack of consistency in the measurement and reporting of drug susceptibility data. In fact, most studies were conducted using a disk diffusion technique similar to this study 53 .
The rate of overall MDR S. aureus documented in this study is similar to most other studies [54][55][56][57][58][59] . This might be due to the fact that S. aurous can present and cause infection to diverse sites of a human body beyond being normal commensal. The major antibiotics prescribed for S. aurous in local area were mostly resisted emphasizing the need for continuous investigation as well as etiologic based therapy. However, in Ethiopia majorities of hospitals were not performing bacteriological culture routinely. Hence, clinician need to look for such type of compiled reports for further consideration of their antibiotics selection.
On the other hand, nearly 88% of K. pneumoniae, 75% E. coli and 70% of P. aeruginosa isolates were resistant to the tested antibiotics. These isolates were mainly identified as hospital-acquired pathogen also they were reported to be community-acquired. The ESKAPE pathogens are the leading cause of nosocomial infections and most of them are MDR, which is one of the greatest challenges in clinical practice amongst critically ill and immunocompromised individuals 15 . Our finding is somehow comparable to previous studies where MDR-ESKAPE reported as predominant isolates 55,60-62 . The resistances acquired to available drugs either by mutation or receipt of foreign genetic materials through the transfer of plasmids and transposons of other microorganisms of the surrounding flora.
In general, the Gram-negative ESKAPE pathogens drug resistance in this study was somehow comparable with others finding for instance, with a study in Latin American countries 63 , Sudan 64 Northern Ethiopia 9 and the systematic reviews conducted in Ethiopia 65,66 . According to the systematic review, the Gram-negative ESKAPE pathogens were highly drug-resistant with estimates between 76 and 87% 66 . Similarly, high MDR Acinetobacter isolates usually reported around 45% of hospitalized patients globally 14 . In Asia-Pacific countries, around 57% of A. baumannii were resistant to all tested drugs 67 . While our finding showed a bit higher, 70% that is similar with Latin American countries report 63 .
Such type of study has an important effect to help the approval of guidelines to the empirical choice of antibiotic therapy. Moreover, from the patient perspective, an appropriate antibiotic resistance profile is essential to reduce mortality and length of hospitalization, the loss of work and family time associated with increased hospitalization time and subsequent recovery, and even the emotional impact of having a resistant infection. In fact, the economic burden of antimicrobial resistance in the prospect of hospital costs, hospital charges including operation costs, the cost of drugs and tests, and other patient care activities and resources used is also a major concern. Hence, such types of studies have a great role for clinicians as well as for policy makers 68 . Limitations. There were certain limitations in this study: First, this retrospective study does not includes all patients visiting our hospital due to some the fact that in our setup culture is not ordered for non-critical patients. For instant most asymptomatic STI and enteric infections by default managed by empirical therapy. Second, we can't compare our findings by dwelling and wards as these variables were not registered in laboratory log book Third, we did not investigate risk factors for infection as well as for drug resistance as it is retrospective data.

Conclusions
According to our investigation, a high rate of bacterial infection (32.6%) was reported and most of the isolates resist the commonly used antibiotics. K. pneumoniae, S. aureus, E. coli were the common isolates in most clinical samples meanwhile CoNS was major bloodstream infection in neonates. The least resistance antibiotics for gram-positive isolate were cefotaxime, meropenem nitrofurantoin and amoxicillin-calvulanic acid. On the other hand, highest levels of resistance were seen for ciprofloxacin, ampicillin and cotrimoxazole. We advise clinicians to see options from such type of compiled data for their routine empirical therapy of bacterial infections.

Data availability
The data that support the findings of this study are available from the corresponding author on upon reasonable request.